1 /* $NetBSD: rtsock.c,v 1.71 2004/05/25 04:33:59 atatat Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * Copyright (c) 1988, 1991, 1993 34 * The Regents of the University of California. All rights reserved. 35 * 36 * Redistribution and use in source and binary forms, with or without 37 * modification, are permitted provided that the following conditions 38 * are met: 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. Neither the name of the University nor the names of its contributors 45 * may be used to endorse or promote products derived from this software 46 * without specific prior written permission. 47 * 48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 58 * SUCH DAMAGE. 59 * 60 * @(#)rtsock.c 8.7 (Berkeley) 10/12/95 61 */ 62 63 #include <sys/cdefs.h> 64 __KERNEL_RCSID(0, "$NetBSD: rtsock.c,v 1.71 2004/05/25 04:33:59 atatat Exp $"); 65 66 #include "opt_inet.h" 67 68 #include <sys/param.h> 69 #include <sys/systm.h> 70 #include <sys/proc.h> 71 #include <sys/mbuf.h> 72 #include <sys/socket.h> 73 #include <sys/socketvar.h> 74 #include <sys/domain.h> 75 #include <sys/protosw.h> 76 #include <sys/sysctl.h> 77 78 #include <net/if.h> 79 #include <net/route.h> 80 #include <net/raw_cb.h> 81 82 #include <machine/stdarg.h> 83 84 extern struct domain routedomain; /* or at least forward */ 85 86 struct sockaddr route_dst = { 2, PF_ROUTE, }; 87 struct sockaddr route_src = { 2, PF_ROUTE, }; 88 struct sockproto route_proto = { PF_ROUTE, }; 89 90 struct walkarg { 91 int w_op; 92 int w_arg; 93 int w_given; 94 int w_needed; 95 caddr_t w_where; 96 int w_tmemsize; 97 int w_tmemneeded; 98 caddr_t w_tmem; 99 }; 100 101 static struct mbuf *rt_msg1(int, struct rt_addrinfo *, caddr_t, int); 102 static int rt_msg2(int, struct rt_addrinfo *, caddr_t, struct walkarg *, int *); 103 static int rt_xaddrs(const char *, const char *, struct rt_addrinfo *); 104 static int sysctl_dumpentry(struct radix_node *, void *); 105 static int sysctl_iflist(int, struct walkarg *, int); 106 static int sysctl_rtable(SYSCTLFN_PROTO); 107 static __inline void rt_adjustcount(int, int); 108 109 /* Sleazy use of local variables throughout file, warning!!!! */ 110 #define dst info.rti_info[RTAX_DST] 111 #define gate info.rti_info[RTAX_GATEWAY] 112 #define netmask info.rti_info[RTAX_NETMASK] 113 #define genmask info.rti_info[RTAX_GENMASK] 114 #define ifpaddr info.rti_info[RTAX_IFP] 115 #define ifaaddr info.rti_info[RTAX_IFA] 116 #define brdaddr info.rti_info[RTAX_BRD] 117 118 static __inline void 119 rt_adjustcount(int af, int cnt) 120 { 121 route_cb.any_count += cnt; 122 switch (af) { 123 case AF_INET: 124 route_cb.ip_count += cnt; 125 return; 126 #ifdef INET6 127 case AF_INET6: 128 route_cb.ip6_count += cnt; 129 return; 130 #endif 131 case AF_IPX: 132 route_cb.ipx_count += cnt; 133 return; 134 case AF_NS: 135 route_cb.ns_count += cnt; 136 return; 137 case AF_ISO: 138 route_cb.iso_count += cnt; 139 return; 140 } 141 } 142 143 /*ARGSUSED*/ 144 int 145 route_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam, 146 struct mbuf *control, struct proc *p) 147 { 148 int error = 0; 149 struct rawcb *rp = sotorawcb(so); 150 int s; 151 152 if (req == PRU_ATTACH) { 153 MALLOC(rp, struct rawcb *, sizeof(*rp), M_PCB, M_WAITOK); 154 if ((so->so_pcb = rp) != NULL) 155 memset(so->so_pcb, 0, sizeof(*rp)); 156 157 } 158 if (req == PRU_DETACH && rp) 159 rt_adjustcount(rp->rcb_proto.sp_protocol, -1); 160 s = splsoftnet(); 161 162 /* 163 * Don't call raw_usrreq() in the attach case, because 164 * we want to allow non-privileged processes to listen on 165 * and send "safe" commands to the routing socket. 166 */ 167 if (req == PRU_ATTACH) { 168 if (p == 0) 169 error = EACCES; 170 else 171 error = raw_attach(so, (int)(long)nam); 172 } else 173 error = raw_usrreq(so, req, m, nam, control, p); 174 175 rp = sotorawcb(so); 176 if (req == PRU_ATTACH && rp) { 177 if (error) { 178 free((caddr_t)rp, M_PCB); 179 splx(s); 180 return (error); 181 } 182 rt_adjustcount(rp->rcb_proto.sp_protocol, 1); 183 rp->rcb_laddr = &route_src; 184 rp->rcb_faddr = &route_dst; 185 soisconnected(so); 186 so->so_options |= SO_USELOOPBACK; 187 } 188 splx(s); 189 return (error); 190 } 191 192 /*ARGSUSED*/ 193 int 194 route_output(struct mbuf *m, ...) 195 { 196 struct rt_msghdr *rtm = 0; 197 struct radix_node *rn = 0; 198 struct rtentry *rt = 0; 199 struct rtentry *saved_nrt = 0; 200 struct radix_node_head *rnh; 201 struct rt_addrinfo info; 202 int len, error = 0; 203 struct ifnet *ifp = 0; 204 struct ifaddr *ifa = 0; 205 struct socket *so; 206 va_list ap; 207 sa_family_t family; 208 209 va_start(ap, m); 210 so = va_arg(ap, struct socket *); 211 va_end(ap); 212 213 #define senderr(e) do { error = e; goto flush;} while (/*CONSTCOND*/ 0) 214 if (m == 0 || ((m->m_len < sizeof(int32_t)) && 215 (m = m_pullup(m, sizeof(int32_t))) == 0)) 216 return (ENOBUFS); 217 if ((m->m_flags & M_PKTHDR) == 0) 218 panic("route_output"); 219 len = m->m_pkthdr.len; 220 if (len < sizeof(*rtm) || 221 len != mtod(m, struct rt_msghdr *)->rtm_msglen) { 222 dst = 0; 223 senderr(EINVAL); 224 } 225 R_Malloc(rtm, struct rt_msghdr *, len); 226 if (rtm == 0) { 227 dst = 0; 228 senderr(ENOBUFS); 229 } 230 m_copydata(m, 0, len, (caddr_t)rtm); 231 if (rtm->rtm_version != RTM_VERSION) { 232 dst = 0; 233 senderr(EPROTONOSUPPORT); 234 } 235 rtm->rtm_pid = curproc->p_pid; 236 memset(&info, 0, sizeof(info)); 237 info.rti_addrs = rtm->rtm_addrs; 238 if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) 239 senderr(EINVAL); 240 info.rti_flags = rtm->rtm_flags; 241 if (dst == 0 || (dst->sa_family >= AF_MAX)) 242 senderr(EINVAL); 243 if (gate != 0 && (gate->sa_family >= AF_MAX)) 244 senderr(EINVAL); 245 if (genmask) { 246 struct radix_node *t; 247 t = rn_addmask((caddr_t)genmask, 0, 1); 248 if (t && genmask->sa_len >= ((struct sockaddr *)t->rn_key)->sa_len && 249 Bcmp((caddr_t *)genmask + 1, (caddr_t *)t->rn_key + 1, 250 ((struct sockaddr *)t->rn_key)->sa_len) - 1) 251 genmask = (struct sockaddr *)(t->rn_key); 252 else 253 senderr(ENOBUFS); 254 } 255 256 /* 257 * Verify that the caller has the appropriate privilege; RTM_GET 258 * is the only operation the non-superuser is allowed. 259 */ 260 if (rtm->rtm_type != RTM_GET && 261 suser(curproc->p_ucred, &curproc->p_acflag) != 0) 262 senderr(EACCES); 263 264 switch (rtm->rtm_type) { 265 266 case RTM_ADD: 267 if (gate == 0) 268 senderr(EINVAL); 269 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt); 270 if (error == 0 && saved_nrt) { 271 rt_setmetrics(rtm->rtm_inits, 272 &rtm->rtm_rmx, &saved_nrt->rt_rmx); 273 saved_nrt->rt_refcnt--; 274 saved_nrt->rt_genmask = genmask; 275 } 276 break; 277 278 case RTM_DELETE: 279 error = rtrequest1(rtm->rtm_type, &info, &saved_nrt); 280 if (error == 0) { 281 (rt = saved_nrt)->rt_refcnt++; 282 goto report; 283 } 284 break; 285 286 case RTM_GET: 287 case RTM_CHANGE: 288 case RTM_LOCK: 289 if ((rnh = rt_tables[dst->sa_family]) == 0) { 290 senderr(EAFNOSUPPORT); 291 } 292 rn = rnh->rnh_lookup(dst, netmask, rnh); 293 if (rn == NULL || (rn->rn_flags & RNF_ROOT) != 0) { 294 senderr(ESRCH); 295 } 296 rt = (struct rtentry *)rn; 297 rt->rt_refcnt++; 298 if (rtm->rtm_type != RTM_GET) {/* XXX: too grotty */ 299 struct radix_node *rn; 300 extern struct radix_node_head *mask_rnhead; 301 302 if (Bcmp(dst, rt_key(rt), dst->sa_len) != 0) 303 senderr(ESRCH); 304 if (netmask && (rn = rn_search(netmask, 305 mask_rnhead->rnh_treetop))) 306 netmask = (struct sockaddr *)rn->rn_key; 307 for (rn = rt->rt_nodes; rn; rn = rn->rn_dupedkey) 308 if (netmask == (struct sockaddr *)rn->rn_mask) 309 break; 310 if (rn == 0) 311 senderr(ETOOMANYREFS); 312 rt = (struct rtentry *)rn; 313 } 314 315 switch (rtm->rtm_type) { 316 case RTM_GET: 317 report: 318 dst = rt_key(rt); 319 gate = rt->rt_gateway; 320 netmask = rt_mask(rt); 321 genmask = rt->rt_genmask; 322 if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { 323 if ((ifp = rt->rt_ifp) != NULL) { 324 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr; 325 ifaaddr = rt->rt_ifa->ifa_addr; 326 if (ifp->if_flags & IFF_POINTOPOINT) 327 brdaddr = rt->rt_ifa->ifa_dstaddr; 328 else 329 brdaddr = 0; 330 rtm->rtm_index = ifp->if_index; 331 } else { 332 ifpaddr = 0; 333 ifaaddr = 0; 334 } 335 } 336 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)0, 337 (struct walkarg *)0, &len); 338 if (len > rtm->rtm_msglen) { 339 struct rt_msghdr *new_rtm; 340 R_Malloc(new_rtm, struct rt_msghdr *, len); 341 if (new_rtm == 0) 342 senderr(ENOBUFS); 343 Bcopy(rtm, new_rtm, rtm->rtm_msglen); 344 Free(rtm); rtm = new_rtm; 345 } 346 (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, 347 (struct walkarg *)0, 0); 348 rtm->rtm_flags = rt->rt_flags; 349 rtm->rtm_rmx = rt->rt_rmx; 350 rtm->rtm_addrs = info.rti_addrs; 351 break; 352 353 case RTM_CHANGE: 354 /* 355 * new gateway could require new ifaddr, ifp; 356 * flags may also be different; ifp may be specified 357 * by ll sockaddr when protocol address is ambiguous 358 */ 359 if ((error = rt_getifa(&info)) != 0) 360 senderr(error); 361 if (gate && rt_setgate(rt, rt_key(rt), gate)) 362 senderr(EDQUOT); 363 /* new gateway could require new ifaddr, ifp; 364 flags may also be different; ifp may be specified 365 by ll sockaddr when protocol address is ambiguous */ 366 if (ifpaddr && (ifa = ifa_ifwithnet(ifpaddr)) && 367 (ifp = ifa->ifa_ifp) && (ifaaddr || gate)) 368 ifa = ifaof_ifpforaddr(ifaaddr ? ifaaddr : gate, 369 ifp); 370 else if ((ifaaddr && (ifa = ifa_ifwithaddr(ifaaddr))) || 371 (gate && (ifa = ifa_ifwithroute(rt->rt_flags, 372 rt_key(rt), gate)))) 373 ifp = ifa->ifa_ifp; 374 if (ifa) { 375 struct ifaddr *oifa = rt->rt_ifa; 376 if (oifa != ifa) { 377 if (oifa && oifa->ifa_rtrequest) 378 oifa->ifa_rtrequest(RTM_DELETE, rt, 379 &info); 380 IFAFREE(rt->rt_ifa); 381 rt->rt_ifa = ifa; 382 IFAREF(rt->rt_ifa); 383 rt->rt_ifp = ifp; 384 } 385 } 386 rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, 387 &rt->rt_rmx); 388 if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) 389 rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); 390 if (genmask) 391 rt->rt_genmask = genmask; 392 /* 393 * Fall into 394 */ 395 case RTM_LOCK: 396 rt->rt_rmx.rmx_locks &= ~(rtm->rtm_inits); 397 rt->rt_rmx.rmx_locks |= 398 (rtm->rtm_inits & rtm->rtm_rmx.rmx_locks); 399 break; 400 } 401 break; 402 403 default: 404 senderr(EOPNOTSUPP); 405 } 406 407 flush: 408 if (rtm) { 409 if (error) 410 rtm->rtm_errno = error; 411 else 412 rtm->rtm_flags |= RTF_DONE; 413 } 414 family = dst ? dst->sa_family : 0; 415 if (rt) 416 rtfree(rt); 417 { 418 struct rawcb *rp = 0; 419 /* 420 * Check to see if we don't want our own messages. 421 */ 422 if ((so->so_options & SO_USELOOPBACK) == 0) { 423 if (route_cb.any_count <= 1) { 424 if (rtm) 425 Free(rtm); 426 m_freem(m); 427 return (error); 428 } 429 /* There is another listener, so construct message */ 430 rp = sotorawcb(so); 431 } 432 if (rtm) { 433 m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); 434 if (m->m_pkthdr.len < rtm->rtm_msglen) { 435 m_freem(m); 436 m = NULL; 437 } else if (m->m_pkthdr.len > rtm->rtm_msglen) 438 m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); 439 Free(rtm); 440 } 441 if (rp) 442 rp->rcb_proto.sp_family = 0; /* Avoid us */ 443 if (family) 444 route_proto.sp_protocol = family; 445 if (m) 446 raw_input(m, &route_proto, &route_src, &route_dst); 447 if (rp) 448 rp->rcb_proto.sp_family = PF_ROUTE; 449 } 450 return (error); 451 } 452 453 void 454 rt_setmetrics(u_long which, const struct rt_metrics *in, struct rt_metrics *out) 455 { 456 #define metric(f, e) if (which & (f)) out->e = in->e; 457 metric(RTV_RPIPE, rmx_recvpipe); 458 metric(RTV_SPIPE, rmx_sendpipe); 459 metric(RTV_SSTHRESH, rmx_ssthresh); 460 metric(RTV_RTT, rmx_rtt); 461 metric(RTV_RTTVAR, rmx_rttvar); 462 metric(RTV_HOPCOUNT, rmx_hopcount); 463 metric(RTV_MTU, rmx_mtu); 464 metric(RTV_EXPIRE, rmx_expire); 465 #undef metric 466 } 467 468 #define ROUNDUP(a) \ 469 ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) 470 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) 471 472 static int 473 rt_xaddrs(const char *cp, const char *cplim, struct rt_addrinfo *rtinfo) 474 { 475 const struct sockaddr *sa = NULL; /* Quell compiler warning */ 476 int i; 477 478 for (i = 0; (i < RTAX_MAX) && (cp < cplim); i++) { 479 if ((rtinfo->rti_addrs & (1 << i)) == 0) 480 continue; 481 rtinfo->rti_info[i] = sa = (struct sockaddr *)cp; 482 ADVANCE(cp, sa); 483 } 484 485 /* Check for extra addresses specified. */ 486 if ((rtinfo->rti_addrs & (~0 << i)) != 0) 487 return (1); 488 /* Check for bad data length. */ 489 if (cp != cplim) { 490 if (i == RTAX_NETMASK + 1 && 491 cp - ROUNDUP(sa->sa_len) + sa->sa_len == cplim) 492 /* 493 * The last sockaddr was netmask. 494 * We accept this for now for the sake of old 495 * binaries or third party softwares. 496 */ 497 ; 498 else 499 return (1); 500 } 501 return (0); 502 } 503 504 static struct mbuf * 505 rt_msg1(int type, struct rt_addrinfo *rtinfo, caddr_t data, int datalen) 506 { 507 struct rt_msghdr *rtm; 508 struct mbuf *m; 509 int i; 510 const struct sockaddr *sa; 511 int len, dlen; 512 513 m = m_gethdr(M_DONTWAIT, MT_DATA); 514 if (m == 0) 515 return (m); 516 MCLAIM(m, &routedomain.dom_mowner); 517 switch (type) { 518 519 case RTM_DELADDR: 520 case RTM_NEWADDR: 521 len = sizeof(struct ifa_msghdr); 522 break; 523 524 #ifdef COMPAT_14 525 case RTM_OIFINFO: 526 len = sizeof(struct if_msghdr14); 527 break; 528 #endif 529 530 case RTM_IFINFO: 531 len = sizeof(struct if_msghdr); 532 break; 533 534 case RTM_IFANNOUNCE: 535 len = sizeof(struct if_announcemsghdr); 536 break; 537 538 default: 539 len = sizeof(struct rt_msghdr); 540 } 541 if (len > MHLEN + MLEN) 542 panic("rt_msg1: message too long"); 543 else if (len > MHLEN) { 544 m->m_next = m_get(M_DONTWAIT, MT_DATA); 545 if (m->m_next == NULL) { 546 m_freem(m); 547 return (NULL); 548 } 549 MCLAIM(m->m_next, m->m_owner); 550 m->m_pkthdr.len = len; 551 m->m_len = MHLEN; 552 m->m_next->m_len = len - MHLEN; 553 } else { 554 m->m_pkthdr.len = m->m_len = len; 555 } 556 m->m_pkthdr.rcvif = 0; 557 m_copyback(m, 0, datalen, data); 558 rtm = mtod(m, struct rt_msghdr *); 559 for (i = 0; i < RTAX_MAX; i++) { 560 if ((sa = rtinfo->rti_info[i]) == NULL) 561 continue; 562 rtinfo->rti_addrs |= (1 << i); 563 dlen = ROUNDUP(sa->sa_len); 564 m_copyback(m, len, dlen, (caddr_t)sa); 565 len += dlen; 566 } 567 if (m->m_pkthdr.len != len) { 568 m_freem(m); 569 return (NULL); 570 } 571 rtm->rtm_msglen = len; 572 rtm->rtm_version = RTM_VERSION; 573 rtm->rtm_type = type; 574 return (m); 575 } 576 577 /* 578 * rt_msg2 579 * 580 * fills 'cp' or 'w'.w_tmem with the routing socket message and 581 * returns the length of the message in 'lenp'. 582 * 583 * if walkarg is 0, cp is expected to be 0 or a buffer large enough to hold 584 * the message 585 * otherwise walkarg's w_needed is updated and if the user buffer is 586 * specified and w_needed indicates space exists the information is copied 587 * into the temp space (w_tmem). w_tmem is [re]allocated if necessary, 588 * if the allocation fails ENOBUFS is returned. 589 */ 590 static int 591 rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w, 592 int *lenp) 593 { 594 int i; 595 int len, dlen, second_time = 0; 596 caddr_t cp0; 597 598 rtinfo->rti_addrs = 0; 599 again: 600 switch (type) { 601 602 case RTM_DELADDR: 603 case RTM_NEWADDR: 604 len = sizeof(struct ifa_msghdr); 605 break; 606 #ifdef COMPAT_14 607 case RTM_OIFINFO: 608 len = sizeof(struct if_msghdr14); 609 break; 610 #endif 611 612 case RTM_IFINFO: 613 len = sizeof(struct if_msghdr); 614 break; 615 616 default: 617 len = sizeof(struct rt_msghdr); 618 } 619 if ((cp0 = cp) != NULL) 620 cp += len; 621 for (i = 0; i < RTAX_MAX; i++) { 622 const struct sockaddr *sa; 623 624 if ((sa = rtinfo->rti_info[i]) == 0) 625 continue; 626 rtinfo->rti_addrs |= (1 << i); 627 dlen = ROUNDUP(sa->sa_len); 628 if (cp) { 629 bcopy(sa, cp, (unsigned)dlen); 630 cp += dlen; 631 } 632 len += dlen; 633 } 634 if (cp == 0 && w != NULL && !second_time) { 635 struct walkarg *rw = w; 636 637 rw->w_needed += len; 638 if (rw->w_needed <= 0 && rw->w_where) { 639 if (rw->w_tmemsize < len) { 640 if (rw->w_tmem) 641 free(rw->w_tmem, M_RTABLE); 642 rw->w_tmem = (caddr_t) malloc(len, M_RTABLE, 643 M_NOWAIT); 644 if (rw->w_tmem) 645 rw->w_tmemsize = len; 646 } 647 if (rw->w_tmem) { 648 cp = rw->w_tmem; 649 second_time = 1; 650 goto again; 651 } else { 652 rw->w_tmemneeded = len; 653 return (ENOBUFS); 654 } 655 } 656 } 657 if (cp) { 658 struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; 659 660 rtm->rtm_version = RTM_VERSION; 661 rtm->rtm_type = type; 662 rtm->rtm_msglen = len; 663 } 664 if (lenp) 665 *lenp = len; 666 return (0); 667 } 668 669 /* 670 * This routine is called to generate a message from the routing 671 * socket indicating that a redirect has occurred, a routing lookup 672 * has failed, or that a protocol has detected timeouts to a particular 673 * destination. 674 */ 675 void 676 rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error) 677 { 678 struct rt_msghdr rtm; 679 struct mbuf *m; 680 const struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; 681 682 if (route_cb.any_count == 0) 683 return; 684 memset(&rtm, 0, sizeof(rtm)); 685 rtm.rtm_flags = RTF_DONE | flags; 686 rtm.rtm_errno = error; 687 m = rt_msg1(type, rtinfo, (caddr_t)&rtm, sizeof(rtm)); 688 if (m == 0) 689 return; 690 mtod(m, struct rt_msghdr *)->rtm_addrs = rtinfo->rti_addrs; 691 route_proto.sp_protocol = sa ? sa->sa_family : 0; 692 raw_input(m, &route_proto, &route_src, &route_dst); 693 } 694 695 /* 696 * This routine is called to generate a message from the routing 697 * socket indicating that the status of a network interface has changed. 698 */ 699 void 700 rt_ifmsg(struct ifnet *ifp) 701 { 702 struct if_msghdr ifm; 703 #ifdef COMPAT_14 704 struct if_msghdr14 oifm; 705 #endif 706 struct mbuf *m; 707 struct rt_addrinfo info; 708 709 if (route_cb.any_count == 0) 710 return; 711 memset(&info, 0, sizeof(info)); 712 memset(&ifm, 0, sizeof(ifm)); 713 ifm.ifm_index = ifp->if_index; 714 ifm.ifm_flags = ifp->if_flags; 715 ifm.ifm_data = ifp->if_data; 716 ifm.ifm_addrs = 0; 717 m = rt_msg1(RTM_IFINFO, &info, (caddr_t)&ifm, sizeof(ifm)); 718 if (m == 0) 719 return; 720 route_proto.sp_protocol = 0; 721 raw_input(m, &route_proto, &route_src, &route_dst); 722 #ifdef COMPAT_14 723 memset(&info, 0, sizeof(info)); 724 memset(&oifm, 0, sizeof(oifm)); 725 oifm.ifm_index = ifp->if_index; 726 oifm.ifm_flags = ifp->if_flags; 727 oifm.ifm_data.ifi_type = ifp->if_data.ifi_type; 728 oifm.ifm_data.ifi_addrlen = ifp->if_data.ifi_addrlen; 729 oifm.ifm_data.ifi_hdrlen = ifp->if_data.ifi_hdrlen; 730 oifm.ifm_data.ifi_mtu = ifp->if_data.ifi_mtu; 731 oifm.ifm_data.ifi_metric = ifp->if_data.ifi_metric; 732 oifm.ifm_data.ifi_baudrate = ifp->if_data.ifi_baudrate; 733 oifm.ifm_data.ifi_ipackets = ifp->if_data.ifi_ipackets; 734 oifm.ifm_data.ifi_ierrors = ifp->if_data.ifi_ierrors; 735 oifm.ifm_data.ifi_opackets = ifp->if_data.ifi_opackets; 736 oifm.ifm_data.ifi_oerrors = ifp->if_data.ifi_oerrors; 737 oifm.ifm_data.ifi_collisions = ifp->if_data.ifi_collisions; 738 oifm.ifm_data.ifi_ibytes = ifp->if_data.ifi_ibytes; 739 oifm.ifm_data.ifi_obytes = ifp->if_data.ifi_obytes; 740 oifm.ifm_data.ifi_imcasts = ifp->if_data.ifi_imcasts; 741 oifm.ifm_data.ifi_omcasts = ifp->if_data.ifi_omcasts; 742 oifm.ifm_data.ifi_iqdrops = ifp->if_data.ifi_iqdrops; 743 oifm.ifm_data.ifi_noproto = ifp->if_data.ifi_noproto; 744 oifm.ifm_data.ifi_lastchange = ifp->if_data.ifi_lastchange; 745 oifm.ifm_addrs = 0; 746 m = rt_msg1(RTM_OIFINFO, &info, (caddr_t)&oifm, sizeof(oifm)); 747 if (m == 0) 748 return; 749 route_proto.sp_protocol = 0; 750 raw_input(m, &route_proto, &route_src, &route_dst); 751 #endif 752 } 753 754 /* 755 * This is called to generate messages from the routing socket 756 * indicating a network interface has had addresses associated with it. 757 * if we ever reverse the logic and replace messages TO the routing 758 * socket indicate a request to configure interfaces, then it will 759 * be unnecessary as the routing socket will automatically generate 760 * copies of it. 761 */ 762 void 763 rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) 764 { 765 struct rt_addrinfo info; 766 struct sockaddr *sa = NULL; 767 int pass; 768 struct mbuf *m = NULL; 769 struct ifnet *ifp = ifa->ifa_ifp; 770 771 if (route_cb.any_count == 0) 772 return; 773 for (pass = 1; pass < 3; pass++) { 774 memset(&info, 0, sizeof(info)); 775 if ((cmd == RTM_ADD && pass == 1) || 776 (cmd == RTM_DELETE && pass == 2)) { 777 struct ifa_msghdr ifam; 778 int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; 779 780 ifaaddr = sa = ifa->ifa_addr; 781 ifpaddr = TAILQ_FIRST(&ifp->if_addrlist)->ifa_addr; 782 netmask = ifa->ifa_netmask; 783 brdaddr = ifa->ifa_dstaddr; 784 memset(&ifam, 0, sizeof(ifam)); 785 ifam.ifam_index = ifp->if_index; 786 ifam.ifam_metric = ifa->ifa_metric; 787 ifam.ifam_flags = ifa->ifa_flags; 788 m = rt_msg1(ncmd, &info, (caddr_t)&ifam, sizeof(ifam)); 789 if (m == NULL) 790 continue; 791 mtod(m, struct ifa_msghdr *)->ifam_addrs = 792 info.rti_addrs; 793 } 794 if ((cmd == RTM_ADD && pass == 2) || 795 (cmd == RTM_DELETE && pass == 1)) { 796 struct rt_msghdr rtm; 797 798 if (rt == 0) 799 continue; 800 netmask = rt_mask(rt); 801 dst = sa = rt_key(rt); 802 gate = rt->rt_gateway; 803 memset(&rtm, 0, sizeof(rtm)); 804 rtm.rtm_index = ifp->if_index; 805 rtm.rtm_flags |= rt->rt_flags; 806 rtm.rtm_errno = error; 807 m = rt_msg1(cmd, &info, (caddr_t)&rtm, sizeof(rtm)); 808 if (m == NULL) 809 continue; 810 mtod(m, struct rt_msghdr *)->rtm_addrs = info.rti_addrs; 811 } 812 route_proto.sp_protocol = sa ? sa->sa_family : 0; 813 raw_input(m, &route_proto, &route_src, &route_dst); 814 } 815 } 816 817 /* 818 * This is called to generate routing socket messages indicating 819 * network interface arrival and departure. 820 */ 821 void 822 rt_ifannouncemsg(struct ifnet *ifp, int what) 823 { 824 struct if_announcemsghdr ifan; 825 struct mbuf *m; 826 struct rt_addrinfo info; 827 828 if (route_cb.any_count == 0) 829 return; 830 memset(&info, 0, sizeof(info)); 831 memset(&ifan, 0, sizeof(ifan)); 832 ifan.ifan_index = ifp->if_index; 833 strlcpy(ifan.ifan_name, ifp->if_xname, sizeof(ifan.ifan_name)); 834 ifan.ifan_what = what; 835 m = rt_msg1(RTM_IFANNOUNCE, &info, (caddr_t)&ifan, sizeof(ifan)); 836 if (m == 0) 837 return; 838 route_proto.sp_protocol = 0; 839 raw_input(m, &route_proto, &route_src, &route_dst); 840 } 841 842 /* 843 * This is used in dumping the kernel table via sysctl(). 844 */ 845 static int 846 sysctl_dumpentry(struct radix_node *rn, void *v) 847 { 848 struct walkarg *w = v; 849 struct rtentry *rt = (struct rtentry *)rn; 850 int error = 0, size; 851 struct rt_addrinfo info; 852 853 if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) 854 return 0; 855 memset(&info, 0, sizeof(info)); 856 dst = rt_key(rt); 857 gate = rt->rt_gateway; 858 netmask = rt_mask(rt); 859 genmask = rt->rt_genmask; 860 if (rt->rt_ifp) { 861 ifpaddr = TAILQ_FIRST(&rt->rt_ifp->if_addrlist)->ifa_addr; 862 ifaaddr = rt->rt_ifa->ifa_addr; 863 if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) 864 brdaddr = rt->rt_ifa->ifa_dstaddr; 865 } 866 if ((error = rt_msg2(RTM_GET, &info, 0, w, &size))) 867 return (error); 868 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 869 struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; 870 871 rtm->rtm_flags = rt->rt_flags; 872 rtm->rtm_use = rt->rt_use; 873 rtm->rtm_rmx = rt->rt_rmx; 874 rtm->rtm_index = rt->rt_ifp->if_index; 875 rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; 876 rtm->rtm_addrs = info.rti_addrs; 877 if ((error = copyout(rtm, w->w_where, size)) != 0) 878 w->w_where = NULL; 879 else 880 w->w_where += size; 881 } 882 return (error); 883 } 884 885 static int 886 sysctl_iflist(int af, struct walkarg *w, int type) 887 { 888 struct ifnet *ifp; 889 struct ifaddr *ifa; 890 struct rt_addrinfo info; 891 int len, error = 0; 892 893 memset(&info, 0, sizeof(info)); 894 TAILQ_FOREACH(ifp, &ifnet, if_list) { 895 if (w->w_arg && w->w_arg != ifp->if_index) 896 continue; 897 ifa = TAILQ_FIRST(&ifp->if_addrlist); 898 ifpaddr = ifa->ifa_addr; 899 switch (type) { 900 case NET_RT_IFLIST: 901 error = 902 rt_msg2(RTM_IFINFO, &info, (caddr_t)0, w, &len); 903 break; 904 #ifdef COMPAT_14 905 case NET_RT_OIFLIST: 906 error = 907 rt_msg2(RTM_OIFINFO, &info, (caddr_t)0, w, &len); 908 break; 909 #endif 910 default: 911 panic("sysctl_iflist(1)"); 912 } 913 if (error) 914 return (error); 915 ifpaddr = 0; 916 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 917 switch (type) { 918 case NET_RT_IFLIST: { 919 struct if_msghdr *ifm; 920 921 ifm = (struct if_msghdr *)w->w_tmem; 922 ifm->ifm_index = ifp->if_index; 923 ifm->ifm_flags = ifp->if_flags; 924 ifm->ifm_data = ifp->if_data; 925 ifm->ifm_addrs = info.rti_addrs; 926 error = copyout(ifm, w->w_where, len); 927 if (error) 928 return (error); 929 w->w_where += len; 930 break; 931 } 932 933 #ifdef COMPAT_14 934 case NET_RT_OIFLIST: { 935 struct if_msghdr14 *ifm; 936 937 ifm = (struct if_msghdr14 *)w->w_tmem; 938 ifm->ifm_index = ifp->if_index; 939 ifm->ifm_flags = ifp->if_flags; 940 ifm->ifm_data.ifi_type = ifp->if_data.ifi_type; 941 ifm->ifm_data.ifi_addrlen = 942 ifp->if_data.ifi_addrlen; 943 ifm->ifm_data.ifi_hdrlen = 944 ifp->if_data.ifi_hdrlen; 945 ifm->ifm_data.ifi_mtu = ifp->if_data.ifi_mtu; 946 ifm->ifm_data.ifi_metric = 947 ifp->if_data.ifi_metric; 948 ifm->ifm_data.ifi_baudrate = 949 ifp->if_data.ifi_baudrate; 950 ifm->ifm_data.ifi_ipackets = 951 ifp->if_data.ifi_ipackets; 952 ifm->ifm_data.ifi_ierrors = 953 ifp->if_data.ifi_ierrors; 954 ifm->ifm_data.ifi_opackets = 955 ifp->if_data.ifi_opackets; 956 ifm->ifm_data.ifi_oerrors = 957 ifp->if_data.ifi_oerrors; 958 ifm->ifm_data.ifi_collisions = 959 ifp->if_data.ifi_collisions; 960 ifm->ifm_data.ifi_ibytes = 961 ifp->if_data.ifi_ibytes; 962 ifm->ifm_data.ifi_obytes = 963 ifp->if_data.ifi_obytes; 964 ifm->ifm_data.ifi_imcasts = 965 ifp->if_data.ifi_imcasts; 966 ifm->ifm_data.ifi_omcasts = 967 ifp->if_data.ifi_omcasts; 968 ifm->ifm_data.ifi_iqdrops = 969 ifp->if_data.ifi_iqdrops; 970 ifm->ifm_data.ifi_noproto = 971 ifp->if_data.ifi_noproto; 972 ifm->ifm_data.ifi_lastchange = 973 ifp->if_data.ifi_lastchange; 974 ifm->ifm_addrs = info.rti_addrs; 975 error = copyout(ifm, w->w_where, len); 976 if (error) 977 return (error); 978 w->w_where += len; 979 break; 980 } 981 #endif 982 default: 983 panic("sysctl_iflist(2)"); 984 } 985 } 986 while ((ifa = TAILQ_NEXT(ifa, ifa_list)) != NULL) { 987 if (af && af != ifa->ifa_addr->sa_family) 988 continue; 989 ifaaddr = ifa->ifa_addr; 990 netmask = ifa->ifa_netmask; 991 brdaddr = ifa->ifa_dstaddr; 992 if ((error = rt_msg2(RTM_NEWADDR, &info, 0, w, &len))) 993 return (error); 994 if (w->w_where && w->w_tmem && w->w_needed <= 0) { 995 struct ifa_msghdr *ifam; 996 997 ifam = (struct ifa_msghdr *)w->w_tmem; 998 ifam->ifam_index = ifa->ifa_ifp->if_index; 999 ifam->ifam_flags = ifa->ifa_flags; 1000 ifam->ifam_metric = ifa->ifa_metric; 1001 ifam->ifam_addrs = info.rti_addrs; 1002 error = copyout(w->w_tmem, w->w_where, len); 1003 if (error) 1004 return (error); 1005 w->w_where += len; 1006 } 1007 } 1008 ifaaddr = netmask = brdaddr = 0; 1009 } 1010 return (0); 1011 } 1012 1013 static int 1014 sysctl_rtable(SYSCTLFN_ARGS) 1015 { 1016 void *where = oldp; 1017 size_t *given = oldlenp; 1018 const void *new = newp; 1019 struct radix_node_head *rnh; 1020 int i, s, error = EINVAL; 1021 u_char af; 1022 struct walkarg w; 1023 1024 if (namelen == 1 && name[0] == CTL_QUERY) 1025 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1026 1027 if (new) 1028 return (EPERM); 1029 if (namelen != 3) 1030 return (EINVAL); 1031 af = name[0]; 1032 w.w_tmemneeded = 0; 1033 w.w_tmemsize = 0; 1034 w.w_tmem = NULL; 1035 again: 1036 /* we may return here if a later [re]alloc of the t_mem buffer fails */ 1037 if (w.w_tmemneeded) { 1038 w.w_tmem = (caddr_t) malloc(w.w_tmemneeded, M_RTABLE, M_WAITOK); 1039 w.w_tmemsize = w.w_tmemneeded; 1040 w.w_tmemneeded = 0; 1041 } 1042 w.w_op = name[1]; 1043 w.w_arg = name[2]; 1044 w.w_given = *given; 1045 w.w_needed = 0 - w.w_given; 1046 w.w_where = where; 1047 1048 s = splsoftnet(); 1049 switch (w.w_op) { 1050 1051 case NET_RT_DUMP: 1052 case NET_RT_FLAGS: 1053 for (i = 1; i <= AF_MAX; i++) 1054 if ((rnh = rt_tables[i]) && (af == 0 || af == i) && 1055 (error = (*rnh->rnh_walktree)(rnh, 1056 sysctl_dumpentry, &w))) 1057 break; 1058 break; 1059 1060 #ifdef COMPAT_14 1061 case NET_RT_OIFLIST: 1062 error = sysctl_iflist(af, &w, w.w_op); 1063 break; 1064 #endif 1065 1066 case NET_RT_IFLIST: 1067 error = sysctl_iflist(af, &w, w.w_op); 1068 } 1069 splx(s); 1070 1071 /* check to see if we couldn't allocate memory with NOWAIT */ 1072 if (error == ENOBUFS && w.w_tmem == 0 && w.w_tmemneeded) 1073 goto again; 1074 1075 if (w.w_tmem) 1076 free(w.w_tmem, M_RTABLE); 1077 w.w_needed += w.w_given; 1078 if (where) { 1079 *given = w.w_where - (caddr_t) where; 1080 if (*given < w.w_needed) 1081 return (ENOMEM); 1082 } else { 1083 *given = (11 * w.w_needed) / 10; 1084 } 1085 return (error); 1086 } 1087 1088 /* 1089 * Definitions of protocols supported in the ROUTE domain. 1090 */ 1091 1092 const struct protosw routesw[] = { 1093 { 1094 SOCK_RAW, &routedomain, 0, PR_ATOMIC|PR_ADDR, 1095 raw_input, route_output, raw_ctlinput, 0, 1096 route_usrreq, 1097 raw_init, 0, 0, 0, 1098 } }; 1099 1100 struct domain routedomain = { 1101 PF_ROUTE, "route", route_init, 0, 0, 1102 routesw, &routesw[sizeof(routesw)/sizeof(routesw[0])] 1103 }; 1104 1105 SYSCTL_SETUP(sysctl_net_route_setup, "sysctl net.route subtree setup") 1106 { 1107 sysctl_createv(clog, 0, NULL, NULL, 1108 CTLFLAG_PERMANENT, 1109 CTLTYPE_NODE, "net", NULL, 1110 NULL, 0, NULL, 0, 1111 CTL_NET, CTL_EOL); 1112 1113 sysctl_createv(clog, 0, NULL, NULL, 1114 CTLFLAG_PERMANENT, 1115 CTLTYPE_NODE, "route", 1116 SYSCTL_DESCR("PF_ROUTE information"), 1117 NULL, 0, NULL, 0, 1118 CTL_NET, PF_ROUTE, CTL_EOL); 1119 sysctl_createv(clog, 0, NULL, NULL, 1120 CTLFLAG_PERMANENT, 1121 CTLTYPE_NODE, "rtable", 1122 SYSCTL_DESCR("Routing table information"), 1123 sysctl_rtable, 0, NULL, 0, 1124 CTL_NET, PF_ROUTE, 0 /* any protocol */, CTL_EOL); 1125 } 1126